Download Network Manipulation (with application to

Network Manipulation (with application to Political issues)
Panagiotis Takis Metaxas
Department of Computer Science
Wellesley College
Wellesley, MA 02481, USA
[email protected]
I. I NTRODUCTION
We live in an increasingly interconnected world, one in
which a growing number of people turn to the web to make
important medical, financial and political decisions [1]. As
more people use the Web’s search engines daily as their
primary source for locating information on many important
issues, search engines are in the position to influence what is
perceived as relevant information through their mechanism
of ranking web pages. However, as studies have shown [2],
interested groups and individuals can also make use of web
spamming mechanisms to trick search engines in ranking
their pages higher than those of their rivals.
The battle for controlling the messages in cyberspace is
spreading over many ideological, cultural, and political issues where controversial positions vie for the public support.
For example, consider issues such as abortion legality and
morality, children vaccination risks, creationism vs. evolution, homosexuality, etc. [3]. Nowhere this battle is more
obvious than when it comes to issues related to national
elections. Obviously, the stakes here are high. If one is able
to influence the public elect officials that will be favorable
to his/her agenda, this will have far-reaching implications.
The term “Web Spam” or Adversarial Web Search [4]
is broadly used to describe misinformation planted on the
Web. Google-bombs are probably the best known examples
of Web spamming, because of their broad coverage in the
press [5]. Web spammers are creating misinformation using
“text spam” and “link spam”, while using “cloaking” to
cover their tracks. This generic categorization of practical
actions does not explain why they are successful, or why
Google Bombs may or may not be be successful.
Recently, spamming techniques have been introduced in
Social Media [6], making it more appropriate to talk about
“Network Manipulation” rather than just Web spam. The
basic techniques of this manipulation, however, can be traced
in the long history of propagandistic techniques in society
[2].
In this presentation proposal we provide an overview of
the technical side of Network Manipulation, and discuss its
connection to Social Propaganda. Further, we describe how
it has been extended in the area of Social Media and discuss
some of its successes and failures when it comes to political
Ranking Techn.
Doc Similarity
Site popularity
Page reputation
Anchor text
Real-time search
Net. manipulation
keyword stuffing
link farms
mutual admiration societies
Google bombs
Twitter bombs
Soc. Propaganda
glittering generalities
bandwagon
testimonials
card stacking
plain folks
Table I
R ANKING TECHNIQUES BY SEARCH ENGINES ARE LISTED ALONG WITH
THE RESPONSE OF THE WEB GRAPH MANIPULATORS AND THEIR
CORRESPONDING PROPAGANDISTIC TECHNIQUES .
issues, especially related to congressional elections. We end
with a discussion of what the search engines have said to
have done so far, and what is likely that they have done
without admitting it.
II. W HAT IS N ETWORK M ANIPULATION
Network Manipulation is the attempt to modify the Web
graph and/or a social network, and thus influence online
network tools in ways beneficial to the manipulators. The
modification of a network is in terms of altering its structure
and/or its contents. The online network tools that manipulators try to influence are typically search engines and online
social media.
One can explain most of the major technological developments in the area of search engine technologies as their
attempt to stop the successes of Web Graph manipulators
(See Table I). For example, Google’s attempt to combat
link farms (groups of interlinked web sites controlled by the
same entity) with the introduction of the famous Page Rank
algorithm was countered by the introduction of “mutual
admiration societies” (organized groups of manipulators who
have achieved high reputation independently for unrelated
themes) [2]. In terms of propagandistic techniques, this
corresponds to “testimonials” often used by advertising
companies: A famous actor playing the doctor on TV urges
the audience to buy a particular pain killer, as if he is an
expert in medicine.
Each of the network manipulation techniques is implemented with altering the structure or the contents of the
network’s components. For example, in creating mutual
admiration societies, the so-called “black hat” search engine
optimization companies organize themselves exchanging
links [7]. To create Google bombs, they announce the terms
to be targetted as anchor text and the links to support,
sometimes even in the open [8], as we discuss next.
Ever since the appearance of the “miserable failure” hoax
that produced search results that included President George
W. Bush (and later, President Barack Obama [9]), Google
bombs have attracted a lot of attention in the media, since
they appear that with little effort, net manipulators can game
the sophisticated algorithms of the search engines. This
practice of “gaming” the search engines is implemented
with mislabeled anchor text techniques (corresponding to the
“card stacking” propagandistic technique; see Appendix??),
in which web site masters and bloggers use the anchor text to
associate an obscure, negative term with a public entity [10].
In particular, during the 2006 US midterm congressional
election, a concerted effort to manipulate ranking results in
order to bring to public attention negative stories about Republican incumbents running for Congress took openly place
under the solicitation of the progressive blog, MyDD.com
(My Direct Democracy) [8].
Can these efforts be stopped? Search engines have tried to
counter this bad publicity by announcing initially a plethora
of features for ranking, and recently more sophisticated
algorithms than Page Rank [11]. These changes seemed to
bear fruits in the 2008 congressional elections where very
few spamming sites rose to the top of search results [12].
It appears, however, that these supposedly “new, sophisticated” (and secret) algorithms would not scale: They are
likely pre-computed search results on a white-list of search
query terms that would likely be bombed.
In particular, these new features and algorithms appeared
to be highly effective in the announced launching of 98
Google bombs [13] during the 2010 congressional elections,
as we predicted [14]. Both in 2008 and 2010, searches
related to congressional candidates would bring up in the top
six results a ranked list of the same sources: one or more
of the campaign sites of the candidates, their official web
pages, their wikipedia page, and google images (Figure 1).
This was consistent across the candidates independently of
their visibility or of the fact that they were under attack
by thousands of political spammers. Moreover, the relative
location of each result in a period of 29 weeks remained
remarkably steady (Figure 2).
At the same time, they proved to be completely ineffective
in at least two examples of network manipulation that
were under their radar screen: The Decor-My-Eyes site [15]
that successfully used bad publicity to rank high, and the
JCPenney case [16] that used anchor text manipulation of
link farms. After these cases became known, their relative
position changed dramatically moving downwards by dozens
of locations per day.
These counter examples provide strong evidence that we
are not talking about new sophisticated algorithms in effect,
Figure 1. Percentage of times a site appeared in a particular position in
the top-10 search results.
Figure 2. Relative change in position of collections of sites during the 29
week period preceding the 2010 congressional elections.
but for old-fashion, hand-crafted list of blacklisted sites and
white-listed terms.
ACKNOWLEDGMENT
The author would like to thank Eni Mustafaraj, Era
Vuksani, Ljubica Ristovska and Dana Bullister for collecting
some of the date presented in this paper. This work was
partially supported by a Brachman-Hoffman Fellowship.
R EFERENCES
[1] Pew Foundation, “Pew internet and american life project.”
http://www.pewinternet.org, 2008.
[2] P. Metaxas, “Web spam, social propaganda and the evolution
of search engine rankings,” Lecture Notes in Bus. Info. Proc.
(LNBIP), accepted; to appear 2010.
[3] M. Hindman, K. Tsioutsiouliklis, and J. Johnson,
“Googlearchy: How a few heavily-linked sites dominate
politics on the web,” in Annual Meeting of the Midwest
Political Science Association, April 3-6 2003.
[4] C. Castillo and B. D. Davison, “Adversarial web search,”
Foundations and Trends in Information Retrieval, vol. 4,
pp. 377–486, June 2010.
[5] T. McNichol, “Your message here.” New York Times, Jan. 22
2004.
[6] C. Grier, K. Thomas, V. Paxson, and M. Zhang, “@spam:
the underground on 140 characters or less,” in Proceedings of
the 17th ACM conference on Computer and communications
security, CCS ’10, (New York, NY, USA), pp. 27–37, ACM,
2010.
III. A PPENDIX : O N P ROPAGANDA T HEORY
Name Calling is the practice of giving an idea a bad label.
It is used to make people reject and condemn the idea
without examining the evidence. For example, using the term
“miserable failure” to refer to political leaders such as US
President George Bush can be thought of as an application
of name calling.
1
• Glittering Generalities is the mirror image of name calling:
Associating an idea with a “virtue word”, in an effort to
make us accept and approve the idea without examining the
evidence. For example, using the term “patriotic” to refer to
illegal actions is a common application of this technique.
• Transfer is the technique by which the propagandist carries
over the authority, sanction, and prestige of something respected and revered to something he would have us accept.
For example, delivering a political speech in a mosque or a
church, or ending a political gathering with a prayer have the
effect of transfer.
• Testimonial is the technique of having some respected person
comment on the quality of an issue on which they have
no qualifications to comment. For example, a famous actor
who plays a medical doctor on a popular TV show tells the
viewers that she only uses a particular pain relief medicine.
The implicit message is that if a famous personality trusts the
medicine, we should too.
• Plain Folks is a technique by which speakers attempt to
convince their audience that they, and their ideas, are “of the
people,” the “plain folks”. For example, politicians sometimes
are seen flipping burgers at a neighborhood diner.
• Card Stacking involves the selection of facts (or falsehoods),
illustrations (or distractions), and logical (or illogical) statements in order to give an incorrect impression. For example,
some activists refer to the Evolution Theory as a theory
teaching that humans came from apes (and not that both apes
and humans have evolved from a common ancestor who was
neither human nor ape).
• Bandwagon is the technique with which the propagandist
attempts to convince us that all members of a group we
belong to accept his ideas and so we should “jump on the
band wagon”. Often, fear is used to reinforce the message.
For example, commercials might show shoppers running to
line up in front of a store before it is open.
The reader should not have much trouble identifying additional
examples of such techniques used in politics or advertising.
We offer here a brief introduction to the theory of propaganda
detection. For more information, see [2].
There are many definitions of propaganda, reflecting its multiple
uses over time. One working definition we will use here is
Propaganda is the attempt to modify human behavior, and thus
influence people’s actions in ways beneficial to propagandists.
Propaganda has a long history in modern society and is often
associated with negative connotation. This was not always the case,
however. The term was first used in 1622, in the establishment
by the Catholic Church of a permanent Sacred Congregation de
Propaganda Fide (for the propagaton of faith), a department which
was trying to spread Catholicism in non-Catholic Countries [17].
Its current meaning comes from the successful Enemy Propaganda
Department in the British Ministry of Information during WWI.
However, it was not until 1938, in the beginning of WWII, that
a theory was developed to detect propagandistic techniques. For
the purposes of this paper we are interested in ways of detecting
propaganda, especially by automatic means.
First developed by the Institute for Propaganda Analysis [18],
classic Propaganda Theory identifies several techniques that propagandists often employ in order to manipulate perception.
1 Name calling and glittering generalities are sometimes referred to as
“word games.”
[7] srainwater,
“Nigritude
ultramarine
http://www.nigritudeultramarines.com/, 2004.
faq.”
[8] T. Zeller Jr., “Gaming the search engine, in a political season.”
New York Times, Nov. 6 2006.
[9] D. Sullivan, “Obama Is “Failure” At Google & “Miserable Failure” At Yahoo.” http://searchengineland.com/yahooobama-is-a-miserable-failure-16286, January 22., 2009.
[10] T. McNichol, “Engineering google results to make a point.”
New York Times, Jan. 22 2004.
[11] S. Hansell, “Google keeps tweaking its search engine.” New
York Times, Jun. 3 2007.
[12] P. Metaxas and E. Mustafaraj, “The battle for the 2008 us
congressional elections on the web,” in Proceedings of the
Web Science 2009 Conference, (Athens, Greece), March 2009.
[13] C.
Bowers,
“Call
to
action:
Googlebombing
the
election.”
Daily
Kos,
http://www.dailykos.com/story/2006/10/22/133437/99,
Last retrieved on Nov. 23, 2010.
[14] S. L. Stirland, “Google is latest weapon vs. gop.” Politico,
http://www.politico.com/news/stories/1010/43767.html, Last
retrieved on Nov. 23, 2010.
[15] D. Segal, “A bully finds a pulpit on the web.” New York
Times, Nov. 20 2010.
[16] D. Segal, “The dirty little secrets of search.” New York Times,
Feb. 12 2011.
[17] D. Welch, “Power of persuasion - propaganda,” History
Today, vol. 49, no. 8, pp. 24–26, 1999.
[18] A. M. Lee and E. B. Lee(eds.), The Fine Art of Propaganda.
The Institute for Propaganda Analysis. Harcourt, Brace and
Co., 1939.
•